Work in-progress.

This is the official repository for an ML Quantitative Precipitation Estimation (ML-QPE) algorithm and accompanying dataset developed in collaboration with the National Weather Service, the University of Oklahoma, the University of Iowa, and the University of North Carolina Chapel Hill.

The following readme introduces the major components of this codebase, including:

1. 2021-2025 MRMS QPE dataset
   • 1M+ samples
   • 220 rain gauge sites
   • HRRR native + derived environmental fields
   • All data aligned to 1km/2min resolution
2. Data analysis/plotting scripts
3. ML-QPE algorithm + benchmarks*

*In development

Download from Hugging Face.

pip install datasets

from datasets import load_dataset

# download the dataset; convert to a dataframe
dataset = load_dataset("leharris3/ccrfcd-mrms-hrrr-env-2021-2025")
df = dataset['train'].to_pandas()

A big thanks to UNITES and L$^3$ at UNC Chapel Hill for lending computing resources to this project.